Ship importance evaluation based on multi-attribute ranking method for maritime safety management

Abstract

As maritime traffic grows, effective ship management is crucial for ensuring safety and optimizing operational efficiency. Traditional approaches to ship importance evaluation often neglect the dynamic interactions and multi-dimensional factors inherent in maritime systems. To address this limitation, a novel framework is introduced that constructs a rule-based complex network from maritime traffic data. Ship importance is then evaluated using a multi-attribute ranking algorithm which integrates five key network metrics: vertex strength, clustering coefficient, degree centrality, betweenness centrality, and closeness centrality. The effectiveness of this approach was validated through network attack comparing it against six single-attribute methods. The results demonstrate the framework's superior performance in identifying critical vessels. Removing the top 50 % of ships ranked by the proposed algorithm caused network efficiency to decrease by 48.4 %. In contrast, removing the same number of ships identified by the best-performing single-attribute method resulted in an efficiency drop of only 39.7 %. This study thus contributes a more robust and effective technique for ship importance evaluation, providing stronger support for decision-making to enhance maritime safety and optimize traffic flow in complex waterways.